Net Election

Mirror, Mirror on the Web

Should you let software decide your vote?

Slate and the Industry Standard join forces to examine the effect of the Internet on Campaign 2000.

At long last, there’s a cure for the intractable problem of the “undecided” voter. Thanks to a new site called the American Presidential Candidate Selector, the Web can actually tell you who you should vote for. Simply type in your opinions on a range of issues, and the site tells you which candidate agrees with you most.

The Candidate Selector is one of several Web tools offered by SelectSmart.com, a company that produces Web sites to help with all sorts of complicated decisions (its slogan is, “Before You Decide”). In addition to the Presidential Candidate Selector, the company offers a Lawn Grass Selector, a Pet Selector, a Hair Products Selector, and a Contraception Selector. Gary Bauer will be matched with your pro-life, creationist views the same way Leave-In Conditioning Spray will be recommended for your dry, frizzy hair.

Here’s how the Candidate Selector works: The site asks your positions on 17 issues, from abortion to defense spending to free trade, and also asks you to choose whether you feel “strongly” or “somewhat strongly” about each of those stands. You can also answer “no preference” to any of the questions. When you’re done, the program spits out a list of all the presidential candidates, ranked and scored in order of how closely their platforms match your views.

The tool claims moderate accuracy. According to a poll on the site, 41 percent of users found that the candidate for whom they intend to vote appeared at the top of their list. Another 19 percent found their top candidate in second place. The poll doesn’t ask the far more interesting question of how many users might change their votes because of the SelectSmart.com results.

I asked a few colleagues with disparate political views to test-drive the Selector. They ended up with funhouse-mirror versions of their actual preferences–accurate in some places, comically distorted in others. The selector placed Bill Bradley and Al Gore, first and second, respectively, on ” Chatterbox” Timothy Noah’s list–wrong order, he says, but not bad. Timothy’s list got less accurate as it continued: It placed Socialist David McReynolds, the Natural Law Party’s John Hagelin, and Warren Beatty (who is not even a candidate) before Timothy’s real-life third choice, John McCain. He observes that “the only accurate score is for Alan Keyes, who I would indeed rank dead last.” Meanwhile, the site told Tucker Carlson, who writes for the conservative Weekly Standard, to vote for Orrin Hatch–“the candidate for whom I have the most contempt,” Tucker says–while awarding John McCain ninth place, behind Hatch, George W. Bush, Harry Browne, Alan Keyes, Steve Forbes, Gary Bauer, Howard Phillips, and Pat Buchanan. “Do they have a drunk guy doing the rankings?” Tucker asks.

Curt Anderson, the site’s designer, sounded sober on the phone, but he has certainly crafted some sloppy questions. One asks users about whether they advocate teaching evolution or creationism in schools. This is not exactly an issue for the next president, but in any case, giving it the same weight as such topics as health care or taxes seems preposterous. Even when Anderson does hit on major issues–crime prevention, environmental protection, foreign policy–his phrasing is without nuance and sometimes without sense. For example, the Selector asks, “Would you prefer your candidate promise to protect or reform (or even abolish) Social Security?” This question is impossible to answer from the choices provided. The distinction between protecting and reforming Social Security is anything but clear: Some candidates such as Steve Forbes suggest that the best way to protect Social Security is to reform it, while even those candidates who advocate preserving the current system–such as Bush and Gore–advocate some significant changes. “Reform” could mean any number of things, from minor tweaks (the addition of individual savings accounts) to more radical fixes (replacing Social Security taxes with privately managed individual accounts). Even if the site asked the question more precisely–say, by asking users if they prefer government-managed savings or private savings–it would reduce an issue of agonizing actuarial detail to a flippant one-liner.

The options for answers are just as clumsy as the questions. The “somewhat” button is ambiguous–if I tell the Selector that I am “somewhat” pro-choice, I might be saying that I want a candidate who is generally pro-choice but who opposes partial-birth abortions. Or I might simply be expressing that abortion is a less urgent issue for me than others. Indeed, the site doesn’t allow you to designate which issues are most important to you. You may care about gun control far more than you do about anything else, but unless you leave the other 16 questions blank, your candidate list won’t reflect your priorities.

The biggest problem with the Presidential Selector, though, is that it reflects an utterly unrealistic sense of how people choose candidates. It undermines itself with its own idealism: Anderson boasts that his Candidate Selector screens out superficial factors such as advertising, hype, and image, and allows voters to base their choices solely on the issues. But few people make up their minds on the basis of positions alone. The site does not try to measure intangibles like integrity, experience, passion, and electability that do–and should–matter to voters.

Right now, the site is most helpful to the minor candidates, whom it gives the same weight as the major players. It helps them find voters–Orrin Hatch’s Web site now links to the Candidate Selector, and Libertarian Harry Browne says that a number of people who found themselves matched to him have become followers.

If the questions and answers were formulated more precisely, the site could make for a fascinating exercise. It would allow you to see how closely you line up with the candidate for whom you intend to vote, and thereby tell you whether your attraction is based on a congruence of views or a more ephemeral sense of goodwill. The technology might usefully be applied in local races, in which it’s hard to sort out the platform of every potential state senator, judge, school board official, and city council member. But for now, the Candidate Selector is no more than a terrific party game–the election-year equivalent of having your palm read.